It's entirely about breaking into the (commodity linux based) control plane for a ASIC miner system.
If they were looking at re-using the actual ASICs for other stuff (do they run microcode? Who knows!) that would be neat. As it is, it's just typical junk-hacking.
------
An ASIC-miner is called an ASIC-miner because it contains ASICs. The author of this post seems to be referring to a system containing ASICs as "an ASIC". This is fundamentally incorrect. Considering the author of the post has also apparently written an entire book about monero, it's a pretty spectacularly egregious error.
The author understands that and is writing for a crypto enthusiast audience that mostly does not understand that. That audience, and even people just tangentially interested, refer to categories of miners as their own terms that are loosely based on the architecture that supports them.
“ASIC” doesn't mean Application Specific Integrated Circuit to them, it means “super fast specialized mining computer that isn't an FPGA and also cant do general computation”
It is a new but decade-old context, so its not worth prefacing that anymore.
Nah. I don’t think its relevant to debate about whether they understand it since they modified the firmware successfully.
The audience they write for doesnt understand “ASIC resistance”, its a meme within that community that has always relied on a misunderstanding of how processors work and is taken advantage of by people that do understand. It would unnecessarily dilute the article to proselytize to the desired audience just because someone that also understands computer architecture might happen to read it.
I would just say you’re late to this one. Like, you would have had an argument in 2012 when there actually was a chance to frame the lingo in the crypto space, but in 2022 for an article from 2019? No.
You should have just sold some hardware or software to miners sometime over the last 10 years.
Again, they poked around in an embedded linux image. There was zero investigation if the ASICs even have firmware, let alone modifying it.
> I would just say you’re late to this one. Like, you would have had an argument in 2012 when there actually was a chance to frame the lingo in the crypto space, but in 2022 for an article from 2019? No.
I mean, ok? If you're writing for a uninformed subset that uses words wrong, and you deliberately use their incorrect terminology, that's kind of worse. I can guarantee that no one who actually works on making the antminer/whatever ASICs calls the entire box "an ASIC".
Writing for a lay audience does not mean deliberately misusing terms.
Meanwhile, the rest of the world uses ASIC to mean Application-Specific Integrated Circuit. Because thats literally what the acronym means.
Is the general purpouse processor on the ASIC or on a separate chip? In the former case the terminology would be ok, in the latter case not, when talking about mucking about the sw on the processor controlling the fixed-function elements.
No, They use a commodity embedded Linux board that sits separate from the boards that contain the ASICs which do the hash calculations. I agree with OP the title is misleading, what this is describing is very pedestrian.
It's entirely about breaking into the (commodity linux based) control plane for a ASIC miner system.
If they were looking at re-using the actual ASICs for other stuff (do they run microcode? Who knows!) that would be neat. As it is, it's just typical junk-hacking.
------
An ASIC-miner is called an ASIC-miner because it contains ASICs. The author of this post seems to be referring to a system containing ASICs as "an ASIC". This is fundamentally incorrect. Considering the author of the post has also apparently written an entire book about monero, it's a pretty spectacularly egregious error.